Septic shock is a major public health problem in both adults and children. Septic shock is a heterogeneous syndrome having highly variable expression in a given patient cohort. A key challenge in the field is to reduce and manage this heterogeneity by more effectively stratifying patients for the purposes of more rational and effective clinical research and individualized clinical management. Over the last 7 years we have developed a genomic expression data base of children with septic shock. We are now proposing to leverage this richly annotated database to develop novel septic shock stratification tools via 3 Specific Aims focused on gene expression-based septic shock subclasses, biomarkers for early detection of septic shock associated renal failure, and genotyping of a novel candidate gene in sepsis biology. In Specific Aim 1 we will derive a classification strategy that groups septic shock patients into distinct subclasses based on a 100 gene expression signature. The classification strategy will be derived in an existing cohort of 180 patients and gene expression measurements will be conducted using the NanoString nCounter platform. Based on our recently published data, we expect that the expression-based subclasses will have clinically important phenotypic differences. The subclass-defining expression signatures will be converted to visually intuitive mosaics using the Gene Expression Dynamic Inspector (GEDI) platform. After generating the subclass- defining mosaics, we will prospectively validate the ability of these mosaics to identify clinically relevant, gene expression-based septic shock subclasses using a separate cohort of 200 patients. In Specific Aim 2 we will derive a risk model for the prediction and early detection of septic shock associated renal failure (SSARF). We have objectively derived a panel of 7, serum-based, candidate protein biomarkers for the early detection of SSARF. We will measure these candidate biomarkers in a derivation cohort of 180 patients and develop a multi-biomarker based risk model. The model will be subsequently validated in a prospectively enrolled cohort of 200 patients. In Specific Aim 3 we will sequence the entire matrix metallopeptidase-8 (MMP-8) gene region and measure associations between sequence variations and illness severity. We will also test associations between MMP-8 sequence variations and expression levels of MMP-8 (mRNA, protein, and activity). The expected deliverables of this application are novel tools for clinically useful stratification of sptic shock.